# Survey questions for VEDLIoT WP2 Architecture Workshop

## General questions
- Which use case do you represent, or do you find most interesting?
	- [ ] Use Case 1: Industry
	- [ ] Use Case 2: Automotive
	- [ ] Use Case 3: Smart Home
	- [ ] All use cases are equally interesting to me.

## Viewpoints of the architecture
- Rank the viewpoints in order of relevance for a deep learning IoT (most relevant is on rank 1, least relevant on rank 8):
	- [ ] Functional viewpoint;
	- [ ] Logical viewpoint;
	- [ ] Technical (physical) viewpoint;
	- [ ] Information / data viewpoint;
	- [ ] Safety and quality;
	- [ ] Security;
	- [ ] Energy;
	- [ ] Cost;
	
## Processing distribution
- From 1-5, where will deep learning processing occur?
	- [ ] Locally on the unit/node
	- [ ] Local central processing unit
	- [ ] Close remote Edge unit
	- [ ] Cloud

## Just-in-time architecture
- Do you believe that it will be easier to develop systems with deep neural networks if the developers can change the architecture when needed (agile architecting)?

## Dynamic aspects of the architecture
- Based on performance criteria (e.g. information throughput), do you think dynamic allocation of processing operations (i.e. inference with deep learning) to different components of the architecture is important?

## Concerns
- What in your opinion is the biggest challenge for an architecture of an AI-intense system? 
